Re: Data truncation problems

Hi folks,

So, I have to admit that I'm not an engineer. Although I speak geek, we're
out of my league now. After discussing the issues w/ the engineer who set up
the server for me and installed the software, we're going to port to a
commercial service licensed during the time in which we set up mod_perl. The
aforementioned engineer is swamped w/ a couple other launches right now, but
I'll ping him in a week or so to get the system details. FWIW, I replicated
the problems using Firefox 1.0.7 on XP Pro.

Thanks for all the help! I'll get those system details.

 - Jennifer

On 3/24/06, Joel Palmius <[email protected]> wrote:
>
> Hrm. I'm not able to replicate the phenomenon. I've copied and pasted
> /etc/services (approx 30k text) into a memo and submitted, which works.
> Then I've copied and pasted it into 10 consecutive memos in one survey
> page, which also works. So with up around 300k data in one page submit,
> I've not had a problem. I've not tested with more data.
>
> Which leaves us with a system-specific problem. Can you post details about
> OS, perl version, mod_perl version, apache version, used browser and any
> other configuration details you think might be interesting? (For example,
> did you do any particular security settings in apache, such as limiting
> caches or similar)
>
>    // Joel
>
> On Tue, 21 Mar 2006, Jennifer Kurkoski wrote:
>
> > Hi Joel & all,
> >
> > Thanks for the reply. Here are a few more details ....
> >
> > The text submission wasn't especially large: just over 2K characters
> > (including spaces). And MEMO never once accepted the whole text block.
> >
> > The total text accepted depended on the point in the survey at which I
> > pasted it in (that is, the question #).
> >
> > The MEMO questions occur near the end and then again at the very end of
> the
> > survey. The later in the survey that I pasted in text blocks, the less
> total
> > text accepted (Note: By "accepted" I mean the amount of text that was
> > present when I pulled the data. There's no error or other indication
> from
> > mod_survey that any problem as occurred).
> >
> > Also of note: once text was truncated in one MEMO question, no text
> appeared
> > at all in any subsequent MEMO questions.
> >
> > The whole survey contains ~26 questions, all on a single page. Some
> > questions are MATRIX (ranging from 2-15 rows), some LIST, some CHOICE,
> some
> > TEXT and some MEMO.
> >
> > I hope that helps. Looking forward to hearing quick and easy solutions.
> ;-)
> >
> > Thanks again,
> >
> > - Jennifer
> > (whose disseration research kinda' depends on this thing)
> >
> >
> >
> > On 3/21/06, Joel Palmius <[email protected]> wrote:
> >>
> >> Never saw this happen with reasonably large posts inside MEMOs myself.
> You
> >> should easily get a few kilobytes of data inside a MEMO. But there are
> >> reasons why large posts could be truncated.
> >>
> >> There are limitations in the http protocol automatically truncating
> very
> >> long data submission when using plain forms. For a GET submit, only
> >> something like 1k data is allowed in a submit, for POST this is larger
> >> although unknown (it is not defined, it is dependent on both browser
> and
> >> server). We currently use POST with no special encoding.
> >>
> >> Further, the MEMO (in practise html's textarea) is limited in length.
> This
> >> limitation is also unknown.
> >>
> >> For very long chunks of data (in the size of hundreds of k) inside one
> >> submit, one would have to use a different form encoding, and probably
> file
> >> uploads.
> >>
> >> Could you post more details on the structure of the survey and how
> large
> >> the posted data was?
> >>
> >>    // Joel
> >>
> >>
> >> On Mon, 20 Mar 2006, Jennifer Kurkoski wrote:
> >>
> >>> Hi all,
> >>>
> >>> I'm having two problems I'm hoping someone here can solve. This is the
> >> first
> >>> (the second coming in a separate msg).
> >>>
> >>> I'm running mod_survey version 3.2.2.
> >>>
> >>> Data are being truncated when written to file. I have several MEMO
> >> blocks in
> >>> my survey. I tested the curious behavior by posting large chunks of
> text
> >> (~3
> >>> long-ish paragraphs) into the MEMO blocks in different configurations
> >>> (skipping Q #1 & 4, for example). The result was always the same: The
> >> first
> >>> instance of the text block would be cut off at some point into it and
> no
> >>> other subsequent text blocks would apppear at all. Sadly, I discovered
> >> this
> >>> bug w/ a live survey. Blah.
> >>>
> >>> I tried searching the archives but found nothing. Anyone have any idea
> >>> what's going on?
> >>>
> >>> Many thanks!!
> >>>
> >>> - Jennifer
> >>>
> >>>
> >>> --
> >>> For best results, follow directions on bottle ...
> >>> and use my permanent address:
> >>>     [email protected]
> >>>
> >> Skickat av Joel Palmius <[email protected]>
> >> till survey-discussion
> >>
> >
> >
> >
> > --
> > For best results, follow directions on bottle ...
> > and use my permanent address:
> >     [email protected]
> >
> Skickat av Joel Palmius <[email protected]>
> till survey-discussion
>



--
For best results, follow directions on bottle ...
and use my permanent address:
     [email protected]